JMIR Formative Research

View query in Help articles search

Field

e.g. mhealth

Operator

Field

Criteria

e.g. John Smith or 0000-0001-6479-5330

Exact Match for Criteria

Yes No

Search Results (1 to 2 of 2 Results)

Download search results: CSV END BibTex RIS

Sort by:

A Machine Learning Approach for Continuous Mining of Nonidentifiable Smartphone Data to Create a Novel Digital Biomarker Detecting Generalized Anxiety Disorder: Prospective Cohort Study

In addition, Tukey post hoc test for pairwise comparison was performed with Cohen d effect size. Detailed results can be found in Multimedia Appendix 2. Multiclass classification accuracy metrics of all algorithms tested in this study (random forest, k-nearest neighbors, logistic regression, XGBoost) using 5-fold cross-validation. a GAD-7: 7-item Generalized Anxiety Disorder Scale. Feature importance of the GAD-7 multiclass XGBoost model. GAD-7: 7-item Generalized Anxiety Disorder Scale.

Soumya Choudhary, Nikita Thomas, Sultan Alshamrani, Girish Srinivasan, Janine Ellenberger, Usman Nawaz, Roy Cohen

JMIR Med Inform 2022;10(8):e38943

Download Citation: END BibTex RIS

Download PDF Download XML

A Machine Learning Approach for Detecting Digital Behavioral Patterns of Depression Using Nonintrusive Smartphone Data (Complementary Path to Patient Health Questionnaire-9 Assessment): Prospective Observational Study

The t test (1-tailed) results of the none versus severe cohorts with P values and Cohen d statistic. Among the classification algorithms, random forest proved to have the highest predictive accuracy (87%). Extreme gradient boosting followed with an accuracy of 86%, whereas the support vector machine classifier had the lowest accuracy (44%), as shown in Table 6.

Soumya Choudhary, Nikita Thomas, Janine Ellenberger, Girish Srinivasan, Roy Cohen

JMIR Form Res 2022;6(5):e37736

Download Citation: END BibTex RIS

Download PDF Download XML

Search Articles

Search Results (1 to 2 of 2 Results)